Anti-sickling effect of MX-1520, a prodrug of vanillin: an in vivo study using rodents

Authors


Dr Toshio Asakura, Division of Hematology, The Children's Hospital of Philadelphia, 3514 Civic Center Blvd, Philadelphia, PA 19104, USA.
E-mail: asakurat@email.chop.edu

Summary

Vanillin, a food additive, covalently binds with sickle haemoglobin (Hb  S), inhibits cell sickling and shifts the oxygen equilibrium curve towards the left. These effects would potentially benefit patients with sickle cell disease (SCD). However, vanillin has no therapeutic effect if given orally because orally administered vanillin is rapidly decomposed in the upper digestive tract. To overcome this problem, a vanillin prodrug, MX-1520, which is biotransformed to vanillin in vivo, was synthesized. Studies using transgenic sickle mice, which nearly exclusively develop pulmonary sequestration upon exposure to hypoxia, showed that oral administration of MX-1520 prior to hypoxia exposure significantly reduced the percentage of sickled cells in the blood. The survival time under severe hypoxic conditions was prolonged from 6·6 ± 0·8 min in untreated animals to 28·8 ± 12 min (P < 0·05) and 31 ± 7·5 min (P < 0·05) for doses of 137·5 and 275 mg/kg respectively. Intraperitoneal injection of MX-1520 to bypass possible degradation in the digestive tract showed that doses as low as 7 mg/kg prolonged the survival time and reduced the percentage of sickled cells during hypoxia exposure. These results demonstrate the potential for MX-1520 to be a new and safe anti-sickling agent for patients with SCD.

Sickle cell disease (SCD) is a hereditary blood disorder caused by a point mutation in the β-globin gene. This mutation results in the substitution of glutamic acid at the β6 position in normal haemoglobin (Hb A) with valine in the sickle haemoglobin (Hb S) (Ingram, 1956; Bunn & Forget, 1986; Schechter et al, 1987; Eaton & Hofrichter, 1990; Embury et al, 1994). This single amino acid substitution in the β-chains dramatically reduces the solubility of the deoxy-form of sickle haemoglobin (deoxy-Hb S). As a result, deoxy-Hb S molecules polymerize inside the red blood cell (RBC), causing deformation of the RBC to assume a sickle shape. Because sickled cells are rigid, they can occlude capillaries and small blood vessels, resulting in ischaemia as well as tissue and organ damage. Although the genetic cause of SCD has been thoroughly elucidated, SCD remains a perplexing disorder for which the exact interrelationship between the gene product (Hb S) and clinical phenotype is the subject of much speculation (Hebbel, 1991). Various types of modulating factors of SCD, including genetic factors, cellular factors and factors that affect the degree of adhesion of sickle erythrocytes to the vascular endothelium, have been reported (Schechter et al, 1987; Hebbel, 1991).

Although numerous anti-sickling agents have been studied (Bunn & Forget, 1986; Schechter et al, 1987; Embury et al, 1994), hydroxyurea (HU) is the only drug that is used clinically for the treatment of SCD. HU prevents erythrocyte sickling by increasing the synthesis of fetal haemoglobin (Charache, 1997; Vichinsky, 1997). Recent studies showed that HU also exerts its beneficial effects in SCD via a number of additional mechanisms, including modification of RBC-endothelial cell interactions, modification of the rheological properties of SS cells, and its myelosuppressive effects, particularly on neutrophils (Hasley & Roberts, 2003). However, not all patients respond favourably to HU and some patients cannot tolerate the drug due to its adverse effects (Goldberg et al, 1992; Platt et al, 1994; Rogers, 1997).

Zaugg et al (1977) incubated sickle erythrocytes (SS cells) with each of 31 different carbonyl compounds and found that aromatic aldehydes, particularly vanillin (2,4-dihydroxybenzaldehyde), a food additive, increased the oxygen affinity of Hb S. They proposed that this group of aldehydes could inhibit sickling because the compounds increased the extent of the oxygen saturation of Hb S under hypoxic conditions. Using high-performance liquid chromatography (HPLC), Abraham et al (1991) showed that vanillin reacted covalently with Hb S, both in solution and in RBCs. They also demonstrated that vanillin shifted the oxygen equilibrium curve (OEC) towards the left and that it inhibited cell sickling using the rheological ektacytometry method.

However, vanillin did not have a significant anti-sickling effect when it was administered orally because it was quickly decomposed in the gastrointestinal (GI) tract (Jenner et al, 1964; Strand & Scheline, 1975; Aroma, 1992; Farthing et al, 1999). To overcome the low oral bioavailability of vanillin, Medinox, Inc. (San Diego, CA, USA) has developed MX-1520, a prodrug of vanillin, which is protected from destruction in the GI tract, yet releases free vanillin to interact with intracellular Hb S in vivo. We report the effects of MX-1520 administered to transgenic (Tg) sickle mice either orally or intraperitoneally.

Materials and methods

Materials and reagents

Vanillin, glucose and bovine serum albumin (BSA) were purchased from Sigma (St Louis, MO, USA). MX-1520 was synthesized by Medinox, Inc.

Animal experiments

All animal experiments were approved by the Animal Care and Use Committee of The Children's Hospital of Philadelphia. Animals were housed and killed in accordance with the guidelines of the National Institutes of Health (1985). The 60% human βS-globin Tg sickle mice were used for this study because the blood oxygen affinity of these mice is similar to that of the blood of patients with SCD and because these mice nearly exclusively develop pulmonary sequestration and die upon exposure to hypoxia (5% oxygen) (Iyamu et al, 2003). Unlike wild-type mice, which have a P50 (the oxygen pressure at which 50% of Hb molecules are saturated with oxygen) of near 40 mmHg (Uchida et al, 1998), the P50 of 60% human βS-globin Tg sickle mice is similar to that of patients with SCD (P50 = 32·0 ± 2·0 mmHg). In addition, the efficacy of a potential anti-sickling agent can be evaluated by assessing the course of the percentage of sickled cells in the blood and the length of the survival time during hypoxic stress, as well as the degree of pulmonary sequestration after exposure to hypoxic stress and death. These mice were produced as described elsewhere (Reilly et al, 1994). The original mice were kindly supplied by Drs Tim Townes (Department of Biochemistry/Molecular Genetics, University of Alabama, Birmingham, AL, USA) and Narla Mohandas (New York Blood Center, New York, NY, USA). As we had observed that the frequency of the development of pulmonary sequestration is inversely related to the oxygen pressure to which the Tg sickle mice are exposed (5–12% oxygen balanced with nitrogen), we exposed Tg sickle mice to 5% oxygen in this study.

Pharmacokinetic study of MX-1520 in Sprague-Dawley rats

The bioavailability of vanillin following administration of MX-1520 or vanillin was studied in Sprague-Dawley rats. A single dose of MX-1520 (186 mg/kg) or vanillin (100 mg/kg) was administered to male Sprague-Dawley rats by oral gavage. These doses represented equimolar amounts of vanillin. Eight animals were included in each group. Aliquots (approximately 200 μl) of blood samples were collected from the cannulated carotid artery at 2, 5, 10, 15, 30 and 45 min, and 1, 1·5, 2 and 3 h after administration of the respective agent. To determine the concentration of vanillin in the blood, aliquots (100 μl) of blood samples were mixed with 400 μl of 8% HClO4 and centrifuged, and 25 μl of the supernatant was analysed by an HPLC apparatus equipped with an ultraviolet detector.

The average plasma concentration of vanillin at each time point was used to calculate the pharmacokinetic parameters for each individual rat. Non-compartmental pharmacokinetic analysis was performed using WinNonlin Professional (version 3.1) software (Pharsight Mountain View, CA, USA). The following parameters were calculated: time of maximum plasma concentration (Tmax), maximum plasma concentration (Cmax), terminal phase half-life (t1/2), area under the curve from zero to the last time point (AUClast), and area under the curve from zero to infinity (AUCinf).

Administration of MX-1520 to Tg sickle mice before exposure to hypoxia

MX-1520 was administered to anaesthetized Tg sickle mice by oral gavage or intraperitoneal (i.p.) injection. For the oral gavage study, Tg sickle mice were divided into three groups. The first group of mice (n = 12) received 275 mg/kg of MX-1520 (MX-1520 was suspended in 0·25% Keltrol), the second group (n = 4) received 137·5 mg/kg of MX-1520, and the control group (n = 8) was given an equivalent volume of 0·25% Keltrol solution. To obtain the maximum blood level, the total amount of MX-1520 or Keltrol solution was divided into three equal portions of 0·1 ml and one portion was administered to the animal at 0, 60 and 120 min by oral gavage. For i.p. administration, MX-1520 was dissolved in 10% dimethyl sulphoxide (DMSO) in saline. The total amount of MX-1520 was divided into three equal portions of 0·1 ml and one portion was given at 0, 20 and 40 min. Doses of 0 (n = 8), 7 (n = 4), 14 (n = 4), 27·5 (n = 5) and 55 mg/kg (n = 3) of MX-1520 were used in the i.p. studies. After the third dose of MX-1520, the mouse was immediately placed in a hypoxia chamber that was flushed with 5% oxygen/95% nitrogen. Mice were exposed to hypoxia for 60 min and the survival time of each mouse under hypoxia was recorded. The study was terminated at 60 min and any surviving mice were killed under anaesthesia at this point.

Blood collection

To study the changes in the percentage of sickled cells in the peripheral blood, blood samples (2–5 μl) were collected under venous oxygen pressure without exposure to air before the hypoxia experiment and approximately every 15 min during hypoxia exposure. The blood was collected from the tail vein and fixed immediately with 2% glutaraldehyde (Asakura et al, 1994). After the mouse died, a blood sample was immediately collected from the vena cava.

Morphological studies of RBCs

Images of RBCs in microslides (Asakura & Mayberry, 1984) were captured using a digital camera, and the morphology of the RBCs was analysed using a computer-assisted image analysis system (Horiuchi et al, 1990, 1996). The percentage of sickled cells and the degree of sickling were determined by measuring parameters that included the surface area, perimeter, and lengths of the long (a) and short axes (b) of each RBC (Horiuchi et al, 1990). Circular shape factor (CSF) represents the degree of deviation from a circle, while elliptical shape factor (ESF) represents the degree of elongation. CSF and ESF were calculated from the following equations: CSF = 4π × area/(perimeter)2, and ESF = b/a. Elongated sickled cells were defined as cells with CSF < 0·8 and ESF < 0·5. Non-elongated sickled cells were defined as those with CSF < 0·8 and ESF > 0·5. The total percentage of sickled cells was calculated as the sum of the percentages of elongated and non-elongated sickled cells. As reported previously (Asakura et al, 1994; Iyamu et al, 2003), exposure of untreated Tg sickle mice to hypoxia increases the percentage of flexible sickled cells with rounded edges from almost 0% to as high as 78%, followed by the formation of rigid sickled cells with spiky edges before death. Pretreatment of Tg sickle mice with an effective drug, however, reduces the percentage of flexible sickled cells with rounded edges to a level less than 40% and inhibits the formation of rigid sickled cells with spiky edges.

Assessment of various organs

When the animals died during or at the end of the experimental period (whichever came first), the lungs, liver, spleen, kidneys and brain were removed and fixed with 10% formalin for histological study. The specimens were sent to the Pathology Department of The Children's Hospital of Philadelphia for sectioning and haematoxylin-eosin staining.

Data analysis

Data are presented as mean ± standard error of the mean (SEM). All data were analysed by paired or unpaired Student's t-test and/or one-way analysis of variance (anova). P < 0·05 was considered to be statistically significant.

Results

Pharmacokinetics of MX-1520 and vanillin

The relative bioavailability of vanillin was studied after oral administration of equimolar amounts of MX-1520 or vanillin in male Sprague-Dawley rats. Rats were utilized instead of mice in this pharmacokinetic comparison because of the practical issues related to obtaining multiple blood samples of adequate volume from a single animal. The bioavailability of vanillin resulting from oral administration of MX-1520 was over 30-fold higher than that resulting from oral administration of vanillin (Table I; Fig 1). The terminal t1/2 time for MX-1520 and vanillin was 45 and 10·2 min respectively. The corresponding vanillin Cmax was nearly fourfold greater upon administration of MX-1520 compared with that of vanillin. These data demonstrated that protected vanillin (prodrug MX-1520) had substantially better vanillin bioavailability properties compared with the unprotected vanillin.

Table I.  Pharmacokinetic parameters of vanillin following oral administration of MX-1520 or vanillin to Sprague-Dawley rats.
 Tmax (min)Cmax (μg/ml)βt1/2 (min)AUClast (min μg/ml)AUCinf (min μg/ml)n
  1. Tmax, the time to maximum concentration of vanillin in the plasma; Cmax, the maximum concentration of vanillin in the plasma; βt1/2, the terminal phase of half-life; AUClast, the area under the curve from zero to the last time point; AUCinf, the area under the curve from zero to infinity.

MX-152010·29·5145·0565·8600·68
Vanillin4·82·4510·217·417·48
Figure 1.

Oral administration of the prodrug MX-1520 results in better vanillin bioavailability compared with oral administration of unprotected vanillin. Equimolar amounts of MX-1520 (•) or vanillin (bsl00001) were administered by oral gavage to male Sprague-Dawley rats as a single dose (n = 8). The concentration of vanillin was determined in blood samples that had been obtained at various time points after dosing and utilized to calculate pharmacokinetic parameters (Table I). Oral administration of the MX-1520 prodrug resulted in superior bioavailability properties compared with oral administration of vanillin.

Oral administration of MX-1520 extended the survival time of Tg sickle mice that were exposed to hypoxia

The effect of oral administration of one of two doses of MX-1520 on the survival time under hypoxia is shown in Fig 2. In the untreated (control) Tg sickle mice, the survival time under the hypoxic condition was 6·6 ± 0·8 min (n = 8). In contrast, the survival times of Tg sickle mice that were treated with 137·5 or 275 mg/kg MX-1520 were 28·8 ± 12·3 min (n = 4, P < 0·05 vs. control) or 31·0 ± 7·5 min (n = 12, P < 0·05 vs. control) respectively.

Figure 2.

Orally administered MX-1520 prolongs the survival time of Tg sickle mice during hypoxia exposure. MX-1520 was administered by oral gavage to Tg sickle mice in three doses at 20-min intervals, to give a total dose of 137·5 or 275 mg/kg. Data are presented as mean ± SEM (n = 8 for the untreated control group, n = 4 for the 137·5 mg/kg group, and n = 12 for the 275 mg/kg group). *P < 0·05, compared with controls.

Oral administration of MX-1520 reduced the percentage of sickled cells in the blood of Tg sickle mice that were exposed to hypoxia

Blood samples were obtained from the Tg sickle mice before and every 15 min during the hypoxia exposure and after death. We determined the percentage of sickled cells in the blood of these Tg sickle mice that had or had not been pretreated with MX-1520 to study whether prolongation of the survival time under hypoxia by MX-1520 was associated with changes in the percentage of sickled cells in the blood. In the control Tg sickle mice, the percentage of sickled cells rapidly rose upon exposure to hypoxia, followed by a sharp decrease immediately before death probably due to the sequestration of sickled cells in the capillaries and small vessels in the lungs (Fig 3). On the contrary, upon hypoxic exposure of Tg sickle mice that had been treated with MX-1520, the percentage of sickled cells in the blood slowly increased over time but never exceeded 30%. The absence of data after approximately 10 min in the control group was due to the deaths of these animals in a relatively short period of time.

Figure 3.

Orally administered MX-1520 reduces sickle cell formation in Tg sickle mice. MX-1520 was administered by oral gavage to Tg sickle mice in three doses at 20-min intervals, to give a total dose of 137·5 or 275 mg/kg. Blood samples were collected from the tail vein before and every 15 min during hypoxia exposure and after the animal died. The percentage of sickled cells was determined by a computer-assisted image analysis system. Data are presented as mean ± SEM [n = 8 for the control group (◆), n = 4 for the 137·5 mg/kg group (▵), and n = 12 for the 275 mg/kg group (□)]. *P < 0·05, compared with controls.

Intraperitoneal administration of MX-1520 extended the survival time of hypoxia-exposed Tg sickle mice at lower drug concentrations

To study whether i.p. administration of MX-1520, which would eliminate the chance of possible degradation of the drug in the digestive tract, is effective, we performed the hypoxia experiment after i.p. administration of MX-1520. As shown in Fig 4, i.p. administration of MX-1520 not only increased the survival time, it did so at a much lower drug dose compared with the dose that was necessary by oral administration. Even at a dose of 7 mg/kg body weight, i.p. administration of MX-1520 significantly prolonged the survival time under hypoxia.

Figure 4.

Intraperitoneal administration of MX-1520 prolongs the survival time of Tg sickle mice during hypoxia exposure. MX-1520 was administered i.p. to Tg sickle mice in three doses at 20-min intervals, to give a total dose of 7, 14, 27·5 or 55 mg/kg. Data are presented as mean ± SEM (n = 8 for control groups (4/4), n = 5 for the 27·5 mg/kg group, n = 3 for the 55 mg/kg group, n = 4 for all other groups). *P < 0·05, compared with the 0 mg/kg and DMSO groups.

Anti-sickling effect of MX-1520 in vivo following i.p. administration

The effect of MX-1520 on sickle cell formation following i.p. administration was also evaluated. The results showed that MX-1520 significantly reduced the percentage of sickled cells in the blood of Tg sickle mice during hypoxic conditions (Fig 5). Without MX-1520 treatment, upon exposure of Tg sickle mice to hypoxia, the percentage of sickled cells rapidly rose to as high as 80% and the animals died. In contrast, the percentage of sickled cells in the blood of MX-1520-treated mice slowly increased over time but rarely exceeded 30%.

Figure 5.

Intraperitoneally administered MX-1520 reduces sickle cell formation in Tg sickle mice. MX-1520 was administered i.p. to Tg sickle mice in three doses at 20-min intervals, to give a total dose of 55 mg/kg (bsl00046), 27·5 mg/kg (◆), or 14 mg/kg (bsl00084). Blood samples were collected from the tail vein before and every 15 min during the hypoxia exposure, and after the animal died. The percentage of sickled cells was determined by a computer-assisted image analysis system as described in the Materials and methods. Data are presented as mean ± SEM [n = 8 for the DMSO group (○), n = 4 for the 14 mg/kg group, n = 5 for the 27·5 mg/kg group, n = 3 for the 55 mg/kg group]. *P < 0·05, compared with the DMSO group. No significant differences were found among the 14, 27·5 and 55 mg/kg groups.

Histopathology of the lungs

The Tg sickle mice used in this study are relatively healthy under normoxic conditions and do not have severe chronic organ damage. However, as reported previously (Iyamu et al, 2003), the control Tg sickle mice that were exposed to severe hypoxia (5% oxygen) died within 15 min. Morphological studies of the RBCs in serial blood samples obtained from these mice showed that flexible, reversibly sickled cells (RSCs) with rounded edges appeared upon exposure of the mice to hypoxia and when these cells started to convert to rigid RSCs with spiky edges, the animal died due to acute severe pulmonary sequestration.

Histological examination of various organs (liver, lungs, spleen, kidneys and brain) of the mice that died during hypoxia exposure showed that the lungs were the only organ that had major pathological changes. It seemed that a massive number of rigid sickled cells had formed in the venous blood and were trapped in the lungs, causing acute, sickling-dependent pulmonary sequestration (Iyamu et al, 2002). The histology of the lungs of representative Tg sickle mice that were pretreated with Keltrol solution or MX-1520 are shown in Fig 6. The lungs of the control mouse showed entrapment of massive numbers of sickled cells in the capillaries and small blood vessels. As a result, the volume of alveolar air space was significantly reduced (Fig 6A and B). On the contrary, the lungs of the Tg sickle mice that were pretreated with MX-1520 and survived for the full 60-min experimental period, showed significantly larger air spaces (Fig 6C and D).

Figure 6.

Histology of the lungs of the hypoxia-exposed Tg sickle mice that were pretreated with MX-1520 or the same volume of keltrol solution. The lungs removed from an untreated Tg sickle mouse that died during hypoxia exposure showed entrapment of massive numbers of sickled cells in the capillaries and small blood vessels due to acute, sickling-dependent pulmonary sequestration (A, B). The lungs of the Tg sickle mice that were pretreated with 275 mg/kg MX-1520 and survived for the full 60-min experimental period showed significantly larger air spaces (C, D). Original magnification: ×25 (A, C); ×250 (B, D).

Discussion

Zaugg et al (1977) reported that vanillin was one of the most effective carbonyl compounds that increase the oxygen affinity of Hb S. Abraham et al (1991) proposed that vanillin exhibits two mechanisms of action for its anti-sickling effect. First, vanillin shifts the OEC towards the left and increases the fraction of oxy-Hb S under hypoxic conditions. Secondly, the Hb S–vanillin complex may stereospecifically prevent the polymerization of deoxy-Hb S. These dual mechanisms of action may have a synergistic effect and vanillin may be more efficacious compared with the combined use of two agents that have one of these two effects. Furthermore, Abraham et al (1991) noted that the vanillin adduct with Hb S was relatively stable. Thus, the anti-sickling effect of vanillin may be long-lasting. Despite this promising activity in vitro, the anti-sickling potential of vanillin as a therapeutic agent for the treatment of patients with SCD has not been realized, probably because vanillin is quickly decomposed in the GI tract (Jenner et al, 1964; Strand & Scheline, 1975; Aroma, 1992; Farthing et al, 1999).

As has been shown by several researchers, aromatic aldehydes that contain a hydroxyl group ortho to aldehyde, including vanillin, form a stable Schiff base with amines, presumably due to the formation of a hydrogen bond with the nitrogen of the Schiff base imine linkage (Zaugg et al, 1977; Beddell et al, 1984). Beddell et al (1984) studied the effect of four different aromatic aldehydes on the oxygen affinity of human haemoglobin and on the sickling of sickle erythrocytes obtained from patients with SCD. They found that all four compounds shifted the OEC of haemoglobin towards the left and inhibited the sickling of sickle erythrocytes under 5% oxygen. Fitzharris et al (1985) reported that intravenous administration of BW12C, a benzyl aldehyde (5-[2-formyl-3-hydroxy]pentanoic acid), to normal healthy volunteers (2–20 mg/kg) caused a dose-dependent left-shift of the blood OEC without significant adverse effects. Keidan et al (1986) studied the effect of BW12C on the oxygen affinity of haemoglobin in eight patients with SCD. A single 1-h intravenous infusion of BW12C showed a dose-dependent increase in whole blood oxygen affinity. At the highest dose given (20 mg/kg body weight), up to 22% of haemoglobin molecules formed haemoglobin adducts without evidence of tissue hypoxia. To identify the binding site of BW12C on haemoglobin molecules, Merrett et al (1986) reacted radio-active BW12C with Hb  A. After treatment with NaBH2, they separated the α- and β-globin chains and found that BW12C bound almost exclusively on the amino group of the terminal amino acid (α1-valine). All of these results showed that benzaldehydes form stable haemoglobin adducts both in vitro and in vivo, and inhibit cell sickling. However, all of the in vivo studies were performed by intravenous administration of the drugs. Naturally, oral administration of any anti-sickling drug is more ideal than other means of administration, including intravenous and i.p. administration.

In contrast to other aldehyde compounds, MX-1520, a prodrug of vanillin, was designed to enhance the oral bioavailability of vanillin. Our pharmacokinetic analysis demonstrated that orally administered MX-1520 resulted in a much higher blood level of vanillin compared to orally administered unprotected vanillin (Table I). Comparison of the bioavailability of vanillin upon oral or i.p. administration of MX-1520 showed that some degradation of MX-1520 occurred in the digestive tract, especially as intraperitoneally administered MX-1520 had a significant in vivo therapeutic effect at much lower concentrations compared to oral doses. Intraperitoneal administration bypasses potential degradation in the digestive tract and processing by the liver. A future dosage form of MX-1520 might therefore include a form of enteric coating to protect the drug during its transit through the stomach and intestine.

Sickled cells in the blood of Tg sickle mice

To determine the percentage of RSCs in the venous blood of Tg sickle mice, blood samples were collected by essentially the same method that was used for blood collection from patients with SCD (Asakura et al, 1994) and Tg sickle mice (Iyamu et al, 2003). Briefly, venous blood samples were collected from the tail vein under venous oxygen pressure without exposure to air, followed by fixing the cells in 2% glutaraldehyde solution that was previously equilibrated with 5%O2/95%N2 (this oxygen pressure corresponds to venous oxygen pressure). The blood samples collected by this method contain two types of RSCs, i.e. ‘flexible RSCs with blunt (rounded) edges’ and ‘rigid RSCs with pointy (spiky) edges’ (Asakura et al, 1994). When Tg sickle mice were exposed to hypoxia, RSCs with rounded edges appeared, followed by the formation of rigid RSCs with spiky edges immediately before death. When untreated Tg sickle mice were exposed to hypoxia, the percentage of flexible RSCs with rounded edges gradually increased to as high as 80%. As these RSCs are flexible, they do not cause vaso-occlusive events. However, after rigid RSCs with spiky edges started to appear, the percentage of circulating sickled cells rapidly decreased and the animal died within 15 min from the onset of hypoxia exposure (Fig 3). The abrupt decrease in the percentage of sickled cells was attributed to the removal of rigid sickled cells from the circulation due to pulmonary sequestration. As shown in Figs 2 and 3, pretreatment of the mice with MX-1520 significantly reduced the percentage of sickled cells in the blood and prolonged the survival time under hypoxic conditions in a dose-dependent manner. The morphological study showed that the sickled cells that formed in the treated mice were mostly flexible RSCs and the number of rigid RSCs was insignificant. These results indicated that MX-1520 inhibits not only the formation of flexible RSCs with rounded edges, but also the conversion of flexible RSCs to rigid sickled cells. We do not know the exact reason why the percentage of sickled cells in the three groups of mice that were pretreated with 14, 27.5 or 55 mg/kg MX-1520 was maintained at the 30% level. We assume that the percentage of flexible RSCs in the blood was affected not only by the anti-sickling drug, but also by other factors, such as the degree of oxygenation of Hb at the lungs, the blood flow rate, the length and diameter of the capillaries, etc. This assumption is supported by our findings that the percentage of RSCs in an individual steady-state patient was maintained in a very narrow range, while the percentage of RSCs among different steady-state patients varied between 4% and 78% (Asakura et al, 1994). We assume that the lowest dose of MX-1520 (14 mg/kg BW) would be sufficient to reduce the percentage of flexible RSCs to the 30% level. The dose-dependent increase in the survival time of Tg sickle mice may be attributed to the dose-dependent inhibition of the conversion of flexible RSCs to rigid sickled cells.

Acknowledgments

We acknowledge Dr Greg Evans at National Heart, Lung and Blood Institute, National Institutes of Health (NIH), for his consistent advice and support in performing this study as the Program Administrator of the Sickle Cell Disease Reference Laboratory. We also acknowledge Emi Asakura for editorial assistance. This work was supported by NIH grant U24 HL 58930, and The Mizuno Fund of The Children's Hospital of Philadelphia.

Declaration of commercial interest: V.V. and M.L. are employed by Medinox, Inc. and have a financial interest in the company.

Note: The GRAS (generally regarded as safe) label does not imply Food and Drug Administration (FDA) approval for the use of vanillin (and certainly not the prodrug under consideration) as an anti-sickling agent, and the safety of the prodrug in animals is currently under study.

Ancillary